Re: RFA: Support Windows extended error numbers in safe_strerror

[Eli Zaretskii <eliz at gnu dot org>]

> Date: Mon, 6 Feb 2006 00:14:51 -0500
> From: Christopher Faylor <cgf-mailing-list-please@cygwin.com>
> 
> >>AFAICT, your objections are, for the most part, not based on fact.
> >
> >They _are_ based on facts (user complaints), you just happen to
> >dismiss those facts as ``irrelevant''.
> 
> You're very tenacious aren't you?

I'm trying very hard not to be tenacious (as in ``stubborn''), but you
cannot expect me to leave remarks like ``not based on fact''
unanswered.  Can't we respectfully disagree without resorting to such
``arguments''?  The fact that we interpret the same facts differently
does not necessarily mean one of us is hallucinating.

> I explained why the issues that you brought up were not problems for a
> cygwin version of gdb.

You asked about incompatibilities between Cygwin and native Windows in
general, so that's what I replied to.  I didn't say anywhere that I
was talking about GDB in the narrow sense.  However, it should be
quite clear that, since Cygwin is a coherent system of tools that need
to work together, any issues with Cygwin at large will necessarily
affect users' will to install and use the Cygwin GDB.  (At least
that's the case with me.)

When I considered installing the Cygwin GDB, the questions that bugged
me were like "will gdbtui work without the Cygwin terminal/rxvt/whatever?",
"will Insight work without Xfree86?", "do I need Bash or other
packages installed, and what are those packages?", etc.  I didn't find
answers to those questions, and AFAICS (although I cannot say I
searched too hard) those answers still aren't available today;
apologies if I missed something obvious.  I ended up installing the
MinGW port becaue it didn't require anything else besides the GDB
binary.  Can you see how this is much easier for someone who likes to
know what is going on on their machines, and doesn't wish to change
the flavor of their system just to use a debugger?

And please don't tell me that Setup does all that automagically for
me: since I have quite an elaborate setup on my machines, I cannot
afford letting automated setup programs download and install packages
without me being in control wrt what is and what isn't installed, to
make sure my configuration doesn't get screwed up.  I don't have the
luxury of spending long hours fixing what some installation broke.

In other words, Cygwin seems to be made for people who are prepared to
marry Cygwin for all their development needs.  It doesn't cater easily
to those, like me, who aren't.  And this problem shows even if all I
want is GDB, to debug my own programs.

> If you want to give details about problems which
> affect the cygwin version of gdb, I'll try to fix them.

I don't think there's anything specific to ``fix'', it's the concept
that is IMHO flawed.  MinGW ports are so much simpler to install
separately, as the need arises, and are so much simpler to control wrt
their overall effect on the system.  Now, if only I could find a
decent MinGW port of a Unixy shell...

Can we now leave this topic alone, please?  I don't expect you to
agree with me or accept my gripes, and neither is this directly
relevant to gdb-patches.  There's no reason to waste our times
disagreeing more, nor flooding others' mailboxes with messages that
don't lead anywhere.